Due to the shortage of liver allografts and the rising prevalence of fatty liver disease in the general population, steatotic liver grafts are considered for transplantation. This condition is an important risk factor for the outcome after transplantation. We here analyze the characteristics of the donor pool offered to the Charité – Universitätsmedizin Berlin from 2010 to 2016 with respect to liver allograft nonacceptance and steatosis hepatis. Of the 2653 organs offered to our center, 19.9% (n=527) were accepted for transplantation, 58.8% (n=1561) were allocated to other centers, and 21.3% (n = 565) were eventually discarded from transplantation. In parallel to an increase of the incidence of steatosis hepatis in the donor pool from 20% in 2010 to 30% in 2016, the acceptance rates for steatotic organs increased in our center from 22.3% to 51.5% in 2016 (p < 0.001), with the majority (86.9%; p > 0.001) having less than 30% macrovesicular steatosis hepatis. However, by 2016, the number of canceled transplantations due to higher grades of steatosis hepatis had significantly increased from 14.7% (n = 15) to 63.6% (42; p < 0.001). The rising prevalence of steatosis hepatis in the donor pool has led to higher acceptance rates of steatotic allografts. Nonetheless, steatosis hepatis remains a predominant phenomenon in discarded organs necessitating future concepts such as organ reconditioning to increase graft utilization.

Normothermic ex vivo liver machine perfusion might be a superior preservation strategy for liver grafts from extended criteria donors. However, standardized small animal models are not available for basic research on machine perfusion of liver grafts. A laboratory‐scaled perfusion system was developed consisting of a custom‐made perfusion chamber, a pressure‐controlled roller pump, and an oxygenator. Male Wistar rat livers were perfused via the portal vein for 6 hours using oxygenated culture medium supplemented with rat erythrocytes. A separate circuit was connected via a dialysis membrane to the main circuit for plasma volume expansion. Glycine was added to the flush solution, the perfusate, and the perfusion circuit. Portal pressure and transaminase release were stable over the perfusion period. Dialysis significantly decreased the potassium concentration of the perfusate and led to significantly higher bile and total urea production. Hematoxylin‐eosin staining and immunostaining for single‐stranded DNA and activated caspase 3 showed less sinusoidal dilatation and tissue damage in livers treated with dialysis and glycine. Although Kupffer cells were preserved, tumor necrosis factor α messenger RNA levels were significantly decreased by both treatments. For proof of concept, the optimized perfusion protocol was tested with donation after circulatory death (DCD) grafts, resulting in significantly lower transaminase release into the perfusate and preserved liver architecture compared with baseline perfusion. In conclusion, our laboratory‐scaled normothermic portovenous ex vivo liver perfusion system enables rat liver preservation for 6 hours. Both dialysis and glycine treatment were shown to be synergistic for preservation of the integrity of normal and DCD liver grafts.

Ex vivo liver machine perfusion (MP) is a promising alternative for preservation of liver grafts from extended criteria donors. Small animal models can be used to evaluate different perfusion conditions. We here describe the development of a miniaturized ex vivo MP system for rat liver grafts, evaluating cell-free and erythrocyte-based perfusion solutions, subnormothermic and normothermic temperatures, and dialysis. A perfusion chamber was designed after a suitable liver position was identified. Normothermic ex vivo liver perfusion (NEVLP) required supplementation of erythrocytes to reduce cell damage. Perfusion with erythrocytes led to rising potassium levels after 12 h (NEVLP, 16.2 mM, interquartile range [IQR]: 5.7 and subnormothermic ex vivo liver perfusion [SNEVLP], 12.8 mM, IQR: 3.5), which were normalized by dialysis using a laboratory dialysis membrane (NEVLP, 6.2 mM, IQR: 0.5 and SNEVLP, 5.3 mM, IQR: 0.1; p = 0.004). Livers treated with NEVLP conditions showed higher bile production (18.52 mg/h/g, IQR: 8.2) compared to livers perfused under SNEVLP conditions (0.4 mg/h/g, IQR: 1.2, p = 0.01). Reducing the perfusion volume from 100 to 50 mL allowed for higher erythrocyte concentrations, leading to significantly lower transaminases (15.75 U/L/mL, IQR: 2.29 vs. 5.97 U/L/mL, IQR: 18.07, p = 0.002). In conclusion, a well-designed perfusion system, appropriate oxygen carriers, dialysis, and miniaturization of the perfusion volume are critical features for successful miniaturized ex vivo liver MP.